Crowdfunding has gained widespread attention in recent years. Despite the huge success of crowdfunding platforms, the percentage of projects that succeed in achieving their desired goal amount is only around 40%. Moreover, many of these crowdfunding platforms follow "all-or-nothing" policy which means the pledged amount is collected only if the goal is reached within a certain predefined time duration. Hence, estimating the probability of success for a project is one of the most important research challenges in the crowdfunding domain. To predict the project success, there is a need for new prediction models that can potentially combine the power of both classification (which incorporate both successful and failed projects) and regression (for estimating the time for success). In this paper, we formulate the project success prediction as a survival analysis problem and apply the censored regression approach where one can perform regression in the presence of partial information. We rigorously study the project success time distribution of crowdfunding data and show that the logistic and loglogistic distributions are a natural choice for learning from such data. We investigate various censored regression models using comprehensive data of 18K Kickstarter (a popular crowdfunding platform) projects and 116K corresponding tweets collected from Twitter. We show that the models that take complete advantage of both the successful and failed projects during the training phase will perform significantly better at predicting the success of future projects compared to the ones that only use the successful projects. We provide a rigorous evaluation on many sets of relevant features and show that adding few temporal features that are obtained at the project's early stages can dramatically improve the performance.

No abstract

Modeling spillover effects from observational data is an important problem in economics, business, and other fields of research. It helps us infer the causality between two seemingly unrelated set of events. For example, if consumer spending in the United States declines, it has spillover effects on economies that depend on the U.S. as their largest export market. In this paper, we aim to infer the causation that results in spillover effects between pairs of entities (or units); we call this effect as paired spillover. To achieve this, we leverage the recent developments in variational inference and deep learning techniques to propose a generative model called Linked Causal Variational Autoencoder (LCVA). Similar to variational autoencoders (VAE), LCVA incorporates an encoder neural network to learn the latent attributes and a decoder network to reconstruct the inputs. However, unlike VAE, LCVA treats the latent attributes as confounders that are assumed to affect both the treatment and the outcome of units. Specifically, given a pair of units u andū, their individual treatment and outcomes, the encoder network of LCVA samples the confounders by conditioning on the observed covariates of u, the treatments of both u andū and the outcome of u. Once inferred, the latent attributes (or confounders) of u captures the spillover effect ofū on u. Using a network of users from job training dataset (LaLonde (1986)) and co-purchase dataset from Amazon e-commerce domain, we show that LCVA is significantly more robust than existing methods in capturing spillover effects. CCS CONCEPTS• Computing methodologies → Supervised learning by regression; • Mathematics of computing → Variational methods; KEYWORDS causal inference; spillover effect; variational autoencoder * Both authors contributed equally to the paper Permission to make digital

Automatic detection of tweets that provide Location-specific information will be extremely useful in conveying geo-location based knowledge to the users. However, there is a significant challenge in retrieving such tweets due to the sparsity of geo-tag information, the short textual nature of tweets, and the lack of pre-defined set of topics. In this paper, we develop a novel framework to identify and summarize tweets that are specific to a location. First, we propose a weighting scheme called Location Centric Word Co-occurrence (LCWC) that uses the content of the tweets and the network information of the twitterers to identify tweets that are location-specific. We evaluate the proposed model using a set of annotated tweets and compare the performance with other weighting schemes studied in the literature. This paper reports three key findings: (a) top trending tweets from a location are poor descriptors of location-specific tweets, (b) ranking tweets purely based on users' geo-location cannot ascertain the location specificity of tweets, and (c) users' network information plays an important role in determining the location-specific characteristics of the tweets. Finally, we train a topic model based on Latent Dirichlet Allocation (LDA) using a large collection of local news database and tweet-based Urls to predict the topics from the location-specific tweets and present them using an interactive web-based interface.

No abstract